In mobile communication, ARQ (Automatic Repeat Request) is applied to downlink data from a radio communication base station apparatus (hereinafter abbreviated to “base station”) to radio communication mobile station apparatuses (hereinafter abbreviated to “mobile stations”). That is, mobile stations feed back response signals representing error detection results of downlink data, to the base station. Mobile stations perform a CRC (Cyclic Redundancy Check) check of downlink data, and, if CRC=OK is found (i.e. if no error is found), feed back an ACK (ACKnowledgement), and, if CRC=NG is found (i.e. if error is found), feed back a NACK (Negative ACKnowledgement), as a response signal to the base station. These response signals are transmitted to the base station using uplink control channels such as a PUCCH (Physical Uplink Control CHannel).
Also, the base station transmits control information for carrying resource allocation results of downlink data, to mobile stations. This control information is transmitted to the mobile stations using downlink control channels such as L1/L2 CCH's (L1/L2 Control CHannels). Each L1/L2 CCH occupies one or a plurality of CCE's (Control Channel Elements) based on the coding rate of control information. For example, when an L1/L2 CCH for carrying control information coded by a rate of 2/3 occupies one CCE, an L1/L2 CCH for carrying control information coded by a rate of 1/3 occupies two CCE's, an L1/L2 CCH for carrying control information coded by a rate of 1/6 occupies four CCE's and an L1/L2 CCH for carrying control information coded by a rate of 1/12 occupies eight CCE's. Also, when one L1/L2 occupies a plurality of CCE's, the CCE's occupied by that one L1/L2 CCH are consecutive. The base station generates an L1/L2 CCH on a per mobile station basis, assigns CCE's to be occupied by L1/L2 CCH's based on the number of CCE's required by control information, and maps the control information on physical resources corresponding to the assigned CCE's and transmits the control information.
Also, studies are underway to map between CCE's and PUCCH's on a one-to-one basis, to use downlink communication resources efficiently without signaling from a base station to mobile stations for reporting the PUCCH's to be used for transmission of response signals (see Non-Patent Document 1). According to this mapping, each mobile station can decide the PUCCH to use to transmit response signals from the mobile station, from the CCE's corresponding to physical resources on which control information for the mobile station is mapped. Therefore, each mobile station maps a response signal from the mobile station on a physical resource, based on the CCE corresponding to a physical resource on which control information directed to the mobile station is mapped. For example, when a CCE corresponding to a physical resource on which control information directed to the mobile station is mapped, is CCE #0, the mobile station decides PUCCH #0 associated with CCE #0 as the PUCCH for the mobile station. Also, for example, when CCE's corresponding to physical resources on which control information directed to the mobile station is mapped are CCE #0 to CCE #3, the mobile station decides PUCCH #0 associated with CCE #0, which is the smallest number in CCE #0 to CCE#3, as the PUCCH for the mobile station, and, when CCE's corresponding to physical resources on which control information directed to the mobile station is mapped are CCE #4 to CCE #7, the mobile station decides PUCCH #4 associated with CCE #4, which is the smallest number in CCE #4 to CCE#7, as the PUCCH for the mobile station.
Also, as shown in FIG. 1, studies are underway to perform code-multiplexing by spreading a plurality of response signals from a plurality of mobile stations using ZAC (Zero Auto Correlation) sequences and Walsh sequences (see Non-Patent Document 2). In FIG. 1, [W0, W1, W2, W3] represents a Walsh sequence with a sequence length of 4. As shown in FIG. 1, in a mobile station, first, a response signal of ACK or NACK is subject to first spreading by a ZAC sequence (with a sequence length of 12) in the frequency domain. Next, the response signal subjected to first spreading is subject to an IFFT (Inverse Fast Fourier Transform) in association with W0 to W3. The response signal spread in the frequency domain by a ZAC sequence with a sequence length of 12 is transformed to a ZAC sequence with a sequence length of 12 in the time domain by this IFFT. Then, the signal subjected to the IFFT is subject to second spreading using a Walsh sequence (with a sequence length of 4). That is, one response signal is allocated to each of four SC-FDMA (Single Carrier-Frequency Division Multiple Access) symbols S0 to S3. Similarly, response signals of other mobile stations are spread using ZAC sequences and Walsh sequences. Here, different mobile stations use ZAC sequences of different cyclic shift values in the time domain (i.e., in the cyclic shift axis) or different Walsh sequences. Here, the sequence length of ZAC sequences in the time domain is 12, so that it is possible to use twelve ZAC sequences of cyclic shift values “0” to “11,” generated from the same ZAC sequence. Also, the sequence length of Walsh sequences is 4, so that it is possible to use four different Walsh sequences. Therefore, in an ideal communication environment, it is possible to code-multiplex maximum forty-eight (12×4) response signals from mobile stations.
Also, as shown in FIG. 1, studies are underway to code-multiplex a plurality of reference signals (e.g. pilot signals) from a plurality of mobile stations (see Non-Patent Document 2). As shown in FIG. 1, in the case of generating three symbols of reference signals R0, R1 and R2, similar to the case of response signals, first, the reference signals are subject to first spreading in the frequency domain by a sequence having characteristics of a ZAC sequence (with a sequence length of 12) in the time domain. Next, the reference signals subjected to first spreading are subject to an IFFT in association with orthogonal sequences with a sequence length of 3, [F0, F1, F2], such as a Fourier sequence. The reference signals spread in the frequency domain are converted by this IFFT to ZAC sequences with a sequence length of 12 in the time domain. Further, these signals subjected to IFFT are subject to second spreading using orthogonal sequences [F0, F1, F2]. That is, one reference signal is allocated to three SC-FDMA symbols R0, R1 and R2. Similarly, other mobile stations allocate one reference signal to three symbols R0, R1 and R2. Here, different mobile stations use ZAC sequences of different cyclic shift values in the time domain or different orthogonal sequences. Here, the sequence length of ZAC sequences in the time domain is 12, so that it is possible to use twelve ZAC sequences of cyclic shift values “0” to “11,” generated from the same ZAC sequence. Also, the sequence length of an orthogonal sequence is 3, so that it is possible to use three different orthogonal sequences. Therefore, in an ideal communication environment, it is possible to code-multiplex maximum thirty-six (12×3) reference signals from mobile stations.
As shown in FIG. 1, seven symbols of S0, S1, R0, R1, R2, S2 and S3 form one slot.
Here, there is substantially no cross correlation between ZAC sequences of different cyclic shift values generated from the same ZAC sequence. Therefore, in an ideal communication environment, a plurality of response signals subjected to spreading and code-multiplexing by ZAC sequences of different cyclic shift values (0 to 11), can be separated in the time domain substantially without inter-code interference, by correlation processing in the base station.
However, due to an influence of, for example, transmission timing difference in mobile stations and multipath delayed waves, a plurality of response signals from a plurality of mobile stations do not always arrive at a base station at the same time. For example, if the transmission timing of a response signal spread by the ZAC sequence of cyclic shift value “0” is delayed from the correct transmission timing, the correlation peak of the ZAC sequence of cyclic shift value “0” may appear in the detection window for the ZAC sequence of cyclic shift value “1.” Further, if a response signal spread by the ZAC sequence of cyclic shift value “0” has a delay wave, an interference leakage due to the delayed wave may appear in the detection window for the ZAC sequence of cyclic shift value “1.” That is, in these cases, the ZAC sequence of cyclic shift value “1” is interfered with by the ZAC sequence of cyclic shift value “0.” On the other hand, if the transmission timing of a response signal spread by the ZAC sequence of cyclic shift value “1” is earlier than the correct transmission timing, the correlation peak of the ZAC sequence of cyclic shift value “1” may appear in the detection window for the ZAC sequence of cyclic shift value “0.” That is, in this case, the ZAC sequence of cyclic shift value “0” is interfered with by the ZAC sequence of cyclic shift value “1.” Therefore, in these cases, the separation performance degrades between a response signal spread by the ZAC sequence of cyclic shift value “0” and a response signal spread by the ZAC sequence of cyclic shift value “1.” That is, if ZAC sequences of adjacent cyclic shift values are used, the separation performance of response signals may degrade.
Therefore, up till now, if a plurality of response signals are code-multiplexed by spreading with ZAC sequences, a sufficient cyclic shift value difference (i.e. cyclic shift interval) is provided between the ZAC sequences, such that inter-code interference is not caused between the ZAC sequences. For example, when the difference between cyclic shift values of ZAC sequences is 2, only six ZAC sequences of cyclic shift values “0,” “2,” “4,” “6,” “8” and “10” or cyclic shift values “1,” “3,” “5,” “7,” “9” and “11” amongst twelve ZAC sequences of cyclic shift values “0” to “12,” are used in first spreading of response signals. Therefore, if a Walsh sequence with a sequence length of 4 is used in second spreading of response signals, it is possible to code-multiplex maximum twenty-four (6×4) response signals from mobile stations.
However, as shown in FIG. 1, the sequence length of an orthogonal sequence used to spread reference signals is 3, and therefore only three different orthogonal sequences can be used to spread reference signals. Consequently, when a plurality of response signals are separated using the reference signals shown in FIG. 1, only maximum eighteen (6×3) response signals from mobile stations can be code-multiplexed. That is, three Walsh sequences are required amongst four Walsh sequences with a sequence length of 4, and therefore one Walsh sequence is not used.
Also, the 1 SC-FDMA symbol shown in FIG. 1 may be referred to as “1 LB (Long Block).” Therefore, a spreading code sequence that is used in spreading in symbol units or LB units, is referred to as a “block-wise spreading code sequence.”
Also, studies are underway to define eighteen PUCCH's as shown in FIG. 2. Normally, the orthogonality of response signals does not collapse between mobile stations using different block-wise spreading code sequences, as long as the mobile stations do not move fast. But, especially if there is a large difference of received power between response signals from a plurality of mobile stations at a base station, one response signal may be interfered with by another response signal between mobile stations using the same block-wise spreading code sequence. For example, in FIG. 2, a response signal using PUCCH #1 (cyclic shift value=2) may be interfered with by a response signal using PUCCH #0 (cyclic shift value=0).
Also, studies are underway to use the constellation shown in FIG. 3 when BPSK is used as the modulation scheme for response signals, and the constellation shown in FIG. 4 when QPSK is used as the modulation scheme for response signals (see Non-Patent Document 3).    Non-Patent Document 1: NTT DoCoMo, Fujitsu, Mitsubishi Electric, “Implicit Resource Allocation of ACK/NACK Signal in E-UTRA Uplink,” Report R1-072439, 3GPP TSG RAN WG1 Meeting #49, Kobe, Japan, May 7-11, 2007.    Non-Patent Document 2: Nokia Siemens Networks, Nokia, “Multiplexing capability of CQIs and ACK/NACKs form different UEs,” Report R1-72315, 3GPP TSG RAN WG1 Meeting #49, Kobe, Japan, May 7-11, 2007.    Non-Patent Document 3: 3GPP TS 36.211 V8.0.0, “Physical Channels and Modulation (Release 8),” September 2007.